Brain Metastases in HER2-Positive Breast Cancer: Current and Novel Treatment Strategies

Simple Summary Development of brain metastases is an important event for patients with breast cancer, and it affects both their survival and their quality of life. Patients with HER2-positive breast cancer are more commonly affected by brain metastases compared to patients with HER2-negative/hormone receptor-positive breast cancer. It is essential to find proper therapies that reduce the risk for metastasis in the brain, as well as agents that are active when metastatic lesions develop. Management of HER2-positive breast cancer has drastically improved in recent years due to the development of several drugs targeting the HER2 receptor. This review aims to provide insight into current and novel treatment strategies for patients with brain metastases from HER2-positive breast cancer. Abstract Development of brain metastases can occur in up to 30–50% of patients with breast cancer, representing a significant impact on an individual patient in terms of survival and quality of life. Patients with HER2-positive breast cancer have an increased risk of developing brain metastases; however, screening for brain metastases is not currently recommended due to the lack of robust evidence to support survival benefit. In recent years, several novel anti-HER2 agents have led to significant improvements in the outcomes of HER2-positive metastatic breast cancer. Despite these advances, brain and leptomeningeal metastases from HER2-positive breast cancer remain a significant cause of morbidity and mortality, and their optimal management remains an unmet need. This review presents an update on the current and novel treatment strategies for patients with brain metastases from HER2-positive breast cancer and discusses the open questions in the field.


Introduction
It is estimated that 30-50% of patients with metastatic breast cancer (MBC) will develop brain metastases (BM) [1,2]. In the Unicancer Epidemiological Strategy and Medical Economics (ESME) MBC database (n = 16,701), 24.6% of the patients developed BM and the risk was higher for patients with HER2-positive/hormone receptor (HR)negative and triple-negative (TNBC) breast cancer [2]. In an individual patient data meta-analysis (n = 9524) by the International Breast Cancer Study Group (IBCSG) including patients with early BC with no prior systemic therapy, the 10-year cumulative incidence of BM was higher among women with HER2-positive disease (6.8% versus 3.5%; p < 0.01) [3]. Other factors associated with BM were an age older than 70 years, the presence of more than two metastatic sites at MBC diagnosis, HR negativity and a more advanced Table 1. Characteristics of studies reporting on outcomes related to brain metastases in patients with HER2-positive breast cancer.

Trastuzumab
Trastuzumab is a monoclonal antibody directed against the extracellular domain (subdomain IV) of HER2 and its discovery has significantly changed the natural history of HER2-positive breast cancer. Despite the transformative impact of trastuzumab in the treatment of patients with early HER2-positive breast cancer, recurrence in the CNS is still a major health issue.
A meta-analysis of randomized trials, including four trials and a total of 9020 patients, showed that adjuvant trastuzumab was associated with an increased incidence of BM as a first site of disease recurrence (1.94% no trastuzumab versus 2.56% trastuzumab, relative risk [RR] 1.35; p = 0.038) [14]. In contrast, in the registHER study, treatment with trastuzumab was associated with significant delay of the time to BM (14 versus 7 months) and also improved survival after BM diagnosis (18 versus 4 months) [6]. Similarly, Dawood and colleagues demonstrated a significantly prolonged time to the development of BM in patients with HER2-positive disease who received trastuzumab compared to those who did not (HR 2.13, 95% CI 1.51-3.00; p < 0.001) [15]. Taken together, these data suggest that, while extracranial disease is effectively controlled with trastuzumab, an intact BBB may prevent trastuzumab penetration to the CNS. However, it can also be hypothesized that the survival benefit achieved by trastuzumab allows for enough time for patients to develop BM, an event not fully prevented by trastuzumab.

Trastuzumab-Pertuzumab
Pertuzumab is a recombinant monoclonal IgG1 antibody directed against the extracellular dimerization domain (subdomain II) of the HER2 receptor. Adding pertuzumab to trastuzumab and paclitaxel in first-line treatment for patients with unresectable and/or metastatic HER2-positive BC significantly increased progression-free survival (PFS) and OS, as demonstrated by the CLEOPATRA phase III trial [49,50]. Patients with clinical and/or radiographic evidence of BM were excluded from the trial. In an exploratory analysis, it was found that the incidence of BM did not significantly differ between the treatment groups (13.7% in the pertuzumab arm versus 12.6% in the placebo arm). However, pertuzumab significantly increased time to the development of BM (15 versus 11.9 months, respectively, HR 0.58; p = 0.0049) and improved survival [16]. The latter can potentially be attributed to the beneficial action of the combination on extracranial disease, intrinsic intracranial activity of pertuzumab or a combination of both.

Ado-Trastuzumab-Emtansine (T-DM1)
T-DM1, an antibody-drug conjugate (ADC) composed of trastuzumab and a derivative of maytansine as cytotoxic payload has been approved by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) in patients with HER2-positive metastatic disease after progression to trastuzumab and a taxane, and in the adjuvant treatment of patients with HER2-positive early breast cancer who have residual invasive disease after neoadjuvant taxane and trastuzumab-based treatment [21,[51][52][53]. In the early setting, the KATHERINE study demonstrated a significant survival benefit provided by adjuvant T-DM1 when administered after failure to achieve pathological complete response (pCR) with neoadjuvant therapy (HR for invasive disease 0.50, 95% CI 0.39-0.64) and the prolongation of time to relapse (invasive disease or death). The incidence of CNS recurrence as a first site of metastasis among patients treated with adjuvant T-DM1 in the study was numerically higher compared to trastuzumab (5.9% versus 4.3%, respectively) [52]. The results underline the efficacy of T-DM1 in this setting and, at the same time, highlight once again the enduring problem with CNS relapse in HER2-positive BC.
In summary, while T-DM1 may have the same limitations as trastuzumab in penetrating the BBB (probably due to large molecular weight), there are some encouraging data regarding its intracranial activity. These findings also raise the question as to whether disruption of the BBB could occur prior to the development of radiological evident CNS metastasis in patients with metastatic HER2-positive BC. This could explain the different efficacy of large molecules in primary and secondary prevention of BM.

Lapatinib
Lapatinib is a reversible dual tyrosine kinase inhibitor (TKI) targeting the epidermal growth factor (EGFR, ErbB-1) and HER2 (ErbB-2), and was initially approved for combination with capecitabine in HER2-positive MBC [54,55]. Several other studies have since investigated lapatinib in the neoadjuvant and metastatic settings in various combinations with other HER2-inhibitors. Given the theoretically higher CNS penetration of a TKI compared to an antibody, lapatinib activity has been explored in patients with CNS disease, with two main objectives: prevent CNS relapse (CEREBEL) [28] and explore its efficacy in patients with CNS metastasis (EGF105084, LANDSCAPE) [29,31].
CEREBEL was a pivotal randomized phase III trial that compared lapatinib and capecitabine with trastuzumab and capecitabine in patients with HER2-positive MBC with no BM at enrollment [28]. The primary endpoint of the trial was incidence of CNS as site of first relapse, based on independent review committee assessment of brain MRI scans. The study was terminated prematurely after the inclusion of 540 patients and demonstrated no significant difference between the groups in terms of incidence of BM (odds ratio [OR] 0.65, 95% CI 0.26-1.63), time to first relapse in the brain (5.7 versus 4.4 months) or progression in the brain at any time (OR 1.14, 95% CI 0.52-2.51) [28]. In the overall population, trastuzumab plus capecitabine demonstrated improved PFS and OS. Although the number and type of prior anti-HER2 regimens may have impacted lapatinib performance, the trial was probably underpowered to detect an existing difference due to its premature termination.

Novel Agents Neratinib
Neratinib is an irreversible pan-HER TKI and has been approved for the management of both early and metastatic BC. The NEfERT-T phase III trial compared neratinib plus paclitaxel versus trastuzumab plus paclitaxel. Patients with BM were eligible if the CNS lesions were asymptomatic [36]. Median PFS in both treatment groups was 12.9 months (HR 1.02, 95% CI 0.81-1.27) but the neratinib-paclitaxel combination led to an increased incidence of grade 3 diarrhea (30.4%) compared to trastuzumab-paclitaxel (3.8%). Despite increased toxicity and lack of PFS benefit, treatment with neratinib plus paclitaxel halved the incidence of CNS recurrence compared to trastuzumab plus paclitaxel (relative risk 0.48, 95% CI 0.29-0.79), and also prolonged the time to development of BM (HR 0.45, 95% CI 0.26-0.78) [36]. Despite fewer patients with BM at baseline in the neratinib-paclitaxel group, the benefit from neratinib remained after adjusting for this difference, regardless of baseline BM status [36]. For the interpretation of these results, it is worth noting that, as screening for synchronous BM at baseline was not performed, the number of metachronous CNS events may have been underestimated.

Trastuzumab
As mentioned earlier, trastuzumab has been generally considered not to cross the BBB due to its high molecular weight. However, in animal models with HER2-positive BM, trastuzumab uptake in the brain and reduction in the BM has been observed [56,57]. Physical disruption of the BBB due to BM growth and increased blood vessel permeability derived from neovessel formation and vascular endothelial growth factor (VEGF) secretion may explain the antibody's brain penetration [58]. However, these observations have not been translated into significant efficacy of single agent trastuzumab in BM in vivo [59,60].
Aiming to improve permeability to the CNS, new mechanisms to disrupt the BBB are being investigated. An example is NEO100, a highly pure derivate of the natural monoterpene perillyl alcohol [61] that has been shown to increase trastuzumab and T-DM1 capacity to penetrate BBB in vitro and in vivo [62]. Intraarterial administration of NEO100 in murine models also facilitated the entry of both drugs into the brain. This was translated into significantly higher apoptotic cell death activity and longer survival for the mice that were treated with the NEO100 combination compared to either anti-HER2 drugs alone or no treatment [62].
Despite a lack of compelling evidence to support trastuzumab monotherapy, some data imply some degree of activity, even for this agent. The already-mentioned registHER trial demonstrated prolonged survival (17.5 versus 3.7 months, HR 0.25, 95% CI 0.20-0.33, p < 0.001) among patients with BM treated with trastuzumab (n = 258), compared to no trastuzumab (n = 119) [6]. A multivariate analysis also identified trastuzumab as an independent favorable prognostic factor for the risk of death (HR 0.33, 95% CI 0.25-0.46, p < 0.001) [6]. Interestingly, radiotherapy was not independently related to improved survival in this cohort. Park et al. also reported survival benefit with trastuzumab in 77 patients with HER2-positive BC with BM [63]. In this retrospective cohort, administration of trastuzumab (n = 42) led to prolonged time to death from BM (14. 9 versus 4 months, p = 0.0005) [63]. Nevertheless, the superiority of a combination of anti-HER2 therapies, rather than a trastuzumab monotherapy, even in extracranial disease, is undisputable and is currently the strategy of choice.

Trastuzumab-Pertuzumab
The phase II PATRICIA trial investigated the efficacy of pertuzumab and high-dose trastuzumab (6 mg/kg weekly) in patients who had experienced progression of BM after CNS radiotherapy but maintained stable extracranial disease. Among the 39 analyzed patients, 11% (95% CI 3-25%) experienced intracranial objective response according to the Response Assessment in Neuro-Oncology Brain Metastases working group (RANO-BM), with a median duration of response (DOR) of 4.6 months [17]. The clinical benefit rate (partial response + complete response + stable disease at 6 months) was 51% [18]. Notably, exploratory studies demonstrated a higher concentration of trastuzumab in serum with this treatment schedule compared to historical controls, suggesting a dose-dependent relationship. Given the limited toxicity reported by the authors, this study provides the possibility for another treatment approach, although more investigations are required to better describe the biological background of the dose-dependency, as well as to validate the toxicity and efficacy results.
Activity on BM was also reported in the RePer study, a retrospective observational study of 264 patients with HER2-positive MBC aiming to explore the efficacy of first-line treatment with pertuzumab/trastuzumab/taxane in a real-world setting. Dual HER2blockade and taxanes resulted in objective response rate (ORR) in 52% of the patients with BM at baseline (n = 21) compared to 82% in those without BM (n = 98). In the whole cohort, 13% of the patients without previous BM experienced progress in the CNS (n = 33) and had worse PFS and two-year OS compared to the patients with BM at base-line; 20 versus 13 months and 77.7% versus 65.6%, respectively [18]. In a similar study, median PFS and OS among 13 patients with BM at baseline treated with first-line trastuzumab/pertuzumab and taxane was 16.8 months and 26.7 months, respectively [19]. Of note, all patients were trastuzumab naïve; thus, this study provides information on the activity on BM of upfront treatment with pertuzumab and trastuzumab.
Benefits from combining the two antibodies with other chemotherapy regimens have also been reported. In the PHEREXA phase III trial, the addition of pertuzumab to capecitabine and trastuzumab significantly improved PFS in the subgroup of patients with BM at baseline (n = 53; HR 0.29, 95% CI 0.15-0.60) [20]. Similarly, although not designed to answer the question of CNS progression, the EORTC 75111-10114 trial enrolled 80 older patients and showed numerical differences in brain-only relapse between treat-ment with pertuzumab and trastuzumab alone (n = 2; 5%), as well as a combination with metronomic oral cyclophosphamide (n = 4; 10%) [64].
In summary, these early studies indicate some activity of the trastuzumab-pertuzumab doublet in BM, always in combination with other agents, and warrant further evaluation.

Ado-Trastuzumab-Emtansine (T-DM1)
The role of T-DM1 in intracranial disease has been assessed in a post hoc analysis of the phase IIIb KAMILLA study [22]. The study enrolled 398 patients with baseline BM, 56.8% of whom had had prior radiotherapy. Among the 126 patients with measurable disease, ORR was 21.4% (95% CI 14.6-29.6) and median PFS and OS were 5.5 (95% CI 5.3-5.6) and 18.9 (95% CI 17.1-21.3) months, respectively. Compared to patients with no BM at baseline, OS was significantly worse among patients with known BM at study enrollment, even after adjustment for other risk factors (adjusted HR 1.18, 95% CI 1.02-1.38) [22]. Interestingly, 67 patients with new brain lesions continued TDM-1 post-progression and had a clinically relevant duration of T-DM1; 8.8 months (0-37 months) and 6.2 months (1-28 months) among patients with and without baseline BM, respectively [22]. Furthermore, in the EMILIA trial (T-DM1 versus lapatinib-capecitabine), no difference in CNS progression rate (22.2% and 16%) and median PFS (5.9 versus 5.7 months) was observed among the 95 patients with previously treated and stable CNS metastases at baseline [21].
In addition, in a smaller observational study (n = 10), objective response was observed in three patients, and median PFS in the CNS was 5 months (95% CI 3.69-6.32 months) [23]. In another study including patients with BM treated with T-DM1 (n = 53), ORR was 25%, PFS was 7 months (95% CI 5.4-8.6 months) and OS was 14 months (95% CI 12.2-15.8) [24]. Finally, another retrospective series (n = 39) also demonstrated clinical benefit in over 50% of the patients and a PFS of 6.1 months (95% CI 5.2-18.3 months), which is in line with the PFS data reported in the KAMILLA study [25].
To summarize, T-DM1 appears to have some activity in the context of BM, with PFS around 5-7 months, although OS remains significantly inferior compared to patients without intracranial disease.

Lapatinib
Lapatinib as monotherapy in patients with progressing BM and prior trastuzumab and radiotherapy was tested in the EGF105084 trial with discouraging results [29]. Only 6% among 242 patients experienced a partial response, and PFS (2.4 months, 1.87-2.79) and OS (6.37, 5.49-8.25) were poor [29]. Interestingly, 20% (n = 10) of the 50 patients that switched to lapatinib plus capecitabine after progression on lapatinib monotherapy achieved a partial response. Likewise, as shown by Blackwell and colleagues, combining lapatinib with trastuzumab led to numerically less progressions of BM compared with lapatinib monotherapy (9/16 and 15/20, respectively), but the numbers are too small for a formal comparison [30]. Similarly, the combination was superior to both lapatinib and trastuzumab monotherapy in a retrospective cohort by Hayashi et al. [65].
As shown above, the lapatinib-capecitabine doublet did not significantly prevent development of BM, but the combination could benefit patients with established, previously untreated BM, as reported in the LANDSCAPE trial [31]. Among 45 enrolled patients, 29 experienced a partial response (65.9%, 95% CI 50.1-79.5%) and median time to progression was 5.5 months, with a significantly longer time to progression among responders (6 months, 95% CI 5.5-7.4) than non-responders (2.8, 95% CI 1.4-4.2). Moreover, in other small cohorts, the doublet yielded a volumetric ORR of 20-38% in patients progressing after prior radiotherapy [32,33]. A pooled analysis performed by Petrelli et al. also demonstrated objective response in about one third of the patients with BM treated with the combination of lapatinib and capecitabine (ORR 29.2%, 95% CI 18.5-42.7) [66]. The ORR was, however, only 18.7% when studies exclusively including patients with progressive BM were analyzed, showing the importance of timely diagnosis and management of BM [66].
In brief, despite the theoretical increased penetrance of a TKI in the CNS, monotherapy with lapatinib has not demonstrated significant intracranial activity. Lapatinib also seems to be more active when given in combination with other agents, such as capecitabine, especially in the context of untreated BM.

Trastuzumab Deruxtecan
Trastuzumab deruxtecan (also known as fam-trastuzumab deruxtecan or DS-8201a) is an ADC composed of a humanized monoclonal antibody anti-human HER2 attached to a topoisomerase I inhibitor payload by a tetrapeptide-based, enzyme-cleavable linker. Compared to T-DM1, trastuzumab deruxtecan has a higher antibody-drug ratio (8 versus 3-4) and is probably more potent than T-DM1 as a result of the properties of its payload that facilitate penetration of the cell membrane of the target or neighboring cells, without requiring high receptor expression [67]. Efficacy of DS-8201a has been assessed in the DESTINY-Breast01 phase II trial that included 184 patients with unresectable or metastatic HER2-positive breast cancer previously treated with T-DM1 [26]. Among this heavily pre-treated population of patients, trastuzumab deruxtecan achieved an ORR of 60.9% and a median PFS of 16.4 months (95% CI 12.7 to not reached) [26].
Patients with BM were eligible for inclusion in the DESTINY Breast01 trial if the brain lesions were previously treated and/or controlled. Among 24 patients with baseline BM, ORR was 58.3% (95% CI 36.6-77.9) and median PFS was 18.1 months (95% CI 6.7-18.1). Progression in the brain was noted in 8% of the patients with baseline BM (n = 2/24) and in 1.3% of the patients without BM at baseline (n = 2/160) [26,27].
Two ongoing studies, NCT04752059 (TUXEDO-1) and NCT04739761 (DESTINY B12), are prospectively enrolling patients to assess the efficacy of trastuzumab deruxtecan in patients with BM. The former is a small phase II study and is already recruiting patients, whereas the latter is a single-arm phase IIIb/4 trial that will include two separate cohorts, distinguishing patients with and without BM, but has not get begun patient enrollment. Both trials will include patients with HER2-positive BC with newly diagnosed or progressing BM and measurable disease according to the RANO-BM criteria. The TUXEDO-1 trial allows for previous treatment with T-DM1, whereas the DESTINY B-12 trial excludes patients with prior treatment with tucatinib. The results of these studies are expected to shed more light on the intracranial activity of trastuzumab deruxtecan.

Neratinib
In the metastatic setting, neratinib was investigated in the phase II TBCRC022 trial in three cohorts: neratinib as a single agent, neratinib as a single agent in patients undergoing surgical excision of CNS lesions and neratinib in combination with capecitabine [34]. In the first cohort, 40 patients were enrolled, 31 of whom had received prior WBRT. The ORR in the CNS was 8% (95% CI, 2% to 22%) and the median PFS was 1.9 months. The third cohort (neratinib plus capecitabine) enrolled 49 patients with HER2-positive BM that had progressed after local treatment in the CNS. Half of the patients in the lapatinib-naïve group (composite CNS ORR 49%, 95% CI 32-66%) and one third among those previously exposed to lapatinib experienced an objective response (composite CNS ORR 33%, 95% CI 10-65%). Median PFS in cohort 3A and 3B was 5.5 months (range 0.8-18.8) and 3.1 months (range 0.7-14.6), respectively, and median OS was 13.3 (range 2.2-27.6) and 15.1 months (range 0.8-23.7), respectively [34]. In spite of some imbalance in the number of previous therapy lines, OS was similar, regardless of prior lapatinib treatment.
Neratinib and capecitabine were compared to lapatinib-capecitabine in the phase III NALA trial. The neratinib combination was not only superior in terms of PFS, but also demonstrated intracranial activity with significantly lower incidence of BM; 22% (95% CI 15.5-30.9%) versus 29.2% in lapatinib-capecitabine (95% CI 22.5-36.1%; HR 0.6, 95% CI 0.6-1.01) [35]. Interestingly, fewer patients in the neratinib arm required intervention for symptomatic CNS metastases (22.8% versus 29.2%, p = 0.043), suggesting an effect of neratinib in the delay of CNS progression, or at least relief of symptoms [35]. Based on these results, neratinib-capecitabine is a preferred combination over lapatinib-capecitabine, although drug availability and reimbursement issues may be an issue.

Tucatinib
Tucatinib is an oral tyrosine kinase inhibitor that is highly selective for the kinase domain of HER2, with minimal inhibition of epidermal growth factor receptor.
Tucatinib has also shown signs of activity in BM in combination with trastuzumab or T-DM1 in phase I studies [38,39]. In a phase I dose-escalation trial of tucatinib with a maximum tolerated dose of 300 mg twice a day (cohort A) or 750 mg once daily (cohort B) in combination with trastuzumab in patients with BM, seven patients in both cohorts together had a clinical benefit rate at 24 weeks [38]. Two of 17 patients in cohort A and one of 17 patients in cohort B experienced partial response, whereas 10 and 16 patients, respectively had stable disease. In another phase I study of tucatinib in combination with T-DM1, 30 of 50 patients treated with 300 mg of tucatinib twice daily had BM at baseline [39]. Median PFS for these patients was 6.7 months (95% CI 4.1-10.2 months) and brain-specific ORR among the 14 patients with measurable BM was 36% [39]. The most effective and/or convenient dose schedule has not been established yet, but the combinations warrant further investigation in, for example, the ongoing NCT04512261 trial.
Tucatinib is, thus far, the first TKI to demonstrate improved antitumor activity against BMs in patients with HER2-positive breast cancer in a randomized controlled trial, and it will probably become an integral part of the management of these patients.

Pyrotinib
Pyrotinib is an oral irreversible inhibitor of the tyrosine kinase activity of EGFR (ErbB-1), HER2 (ErbB-2) and HER4 (ErbB4) [69]. The combination of pyrotinib and capecitabine in the PHENIX phase III trial significantly increased PFS from 4.2 months with single agent capecitabine to 6.9 months in the combination group among the 31 study participants with BM [40].
In a subgroup of 31 patients with BM from real world data reporting on pyrotinib monotherapy or in combinations (n = 122), PFS was 6.7 months (both intra-and extracranial lesions considered) and ORR was 28% [41]. Nevertheless, concomitant local treatment (radiotherapy or surgery) improved response, as seen by the difference in ORR in patients with or without local interventions; 66.6% (n = 9) and 6.3% (n = 16), respectively. Of note, three patients in the first subgroup (33.3%) treated with pyrotinib and capecitabine and concomitant radiotherapy achieved a complete response [41].
Early results from pyrotinib studies are encouraging, and it remains to be seen whether pyrotinib will be added to the armory of anti-HER2 agents.

Margetuximab
Margetuximab is a chimeric IgG1 monoclonal antibody that holds the same epitope specificity of trastuzumab. However, its fragment crystallizable (Fc) domain has been modified to improve affinity for CD16A and reduce affinity for CD32B, resulting in better ADCC activity over trastuzumab in vitro [70].
The antibody has been investigated in the phase III SOPHIA trial that compared margetuximab to trastuzumab, both in combination with chemotherapy (investigator's choice: capecitabine, eribulin, gemcitabine, or vinorelbine) [71]. Patients were eligible if they had progressed after at least two prior HER2-targeted therapies. Patients with BM were allowed if the metastases were treated and stable. A total of 536 patients were enrolled, and 266 (including 37 patients with BM) and 270 patients (including 34 patients with BM) were randomized to margetuximab and trastuzumab combinations, respectively. The margetuximab group demonstrated significantly improved PFS from 4.9 to 5.8 months (HR 0.76, 95% CI 0.59-0.98; p = 0.03). However, the clinical relevance of this improvement of less than a month is debatable. No statistically significant difference was observed in terms of OS (21.6 versus 19.8 months in the margetuximab and trastuzumab groups, respectively), although the final OS analysis is pending [71].
No BM specific endpoints have been reported so far; therefore, it remains unknown whether margetuximab is effective in BM for HER2-positive breast cancer.

Combination of Systemic Therapy with Radiotherapy
Although the anti-HER2 tyrosine kinase inhibitors (TKIs) appear to offer better intracranial disease control compared to the antibodies, local interventions will probably continue to be required for local control and symptom relief. Data informing on the efficacy and safety of anti-HER2 therapies plus radiotherapy (whole brain radiotherapy [WBRT], stereotactic external beam radiotherapy [SBRT] or stereotactic radiosurgery [SRS]) are scarce.

Trastuzumab
Trastuzumab's combination with WBRT in a small cohort of 31 patients appeared to be safe and effective, with an intracranial ORR of 74.2% and a median time to BM progression of 10 months [42]. No comparative results to either trastuzumab or WBRT alone are available; thus, it remains unclear whether the combination is indeed more beneficial than any of the interventions alone.

T-DM1
Among 12 patients receiving T-DM1 either sequentially to SRS (n = 8) or concomitantly with WBRT (n = 4) at Institut Curie, ORR was 50% in both groups and median PFS was 5.5 and 12.5 months, respectively. However, radiation necrosis was a common adverse event and it was observed in 30.8% of the patients [43]. This finding is in concordance with an independent cohort of 23 patients treated with T-DM1 in combination with SRS concurrently (n = 16) or sequentially (n = 7) [44]. Radiation necrosis was then observed in 39.1% of patients compared with 2.5% of 22 patients who did not receive T-DM1. Hence, T-DM1 significantly increased the risk of radiation necrosis by 13.5 times, with a median time from radiotherapy to symptom debut of 16 months (range 1-79) [44]. Preclinical data suggest that DM1 bystander effect over reactive astrocytes enhances Aquaporin 4 expression in previously irradiated human astrocyte cell cultures. Furthermore, T-DM1 also increased radiation-related astrocytic cell death. These results were not seen with trastuzumab, suggesting this adverse event is associated with T-DM1 [44].

Lapatinib
The safety and efficacy of combining anti-HER2 therapies with radiotherapy have been addressed in a phase I trial with lapatinib [45]. A total of 35 women with HER-positive MBC with BM (not exclusively) were enrolled, 28 of them with intracranial measurable disease. The combination demonstrated ORR of 79% (59-92%), much higher than the limited response rate that was observed in lapatinib monotherapy [29,45]. Median PFS was also longer when lapatinib was combined with radiotherapy; 4.8 months compared to the previously mentioned 2.4 months in lapatinib monotherapy [29,45]. However, toxicity was a major limiting factor in this cohort [45], even though main toxicities, such as diarrhea, acneiform rash and asthenia, were attributed to lapatinib. No additional toxicity emanating from the combination was observed.
Local disease control with the combination of lapatinib and SRS was investigated retrospectively in a cohort of 126 patients with BM, 47 of whom received concurrent or sequential lapatinib [46]. The combination significantly decreased local failure and increased OS, with greater benefit in local control among patients with smaller lesions. However, data regarding intracranial response rate were not reported. Concurrent or sequential lapatinib was found to be safe and did not significantly increase cumulative incidence of radiation necrosis [46]. In another retrospective cohort study from the same center, lapatinib in combination with SRS (n = 18) significantly increased intracranial complete response rate (35% versus 11%), without significant benefit in terms of ORR (75% versus 57%) or brain target lesion-progression rate (25% versus 43%) compared to SRS alone (n = 66) [47]. Furthermore, in a phase II study of lapatinib in combination with WBRT reported by Christodoulou et al. (n = 21 breast cancer patients), HER2-positive status (n = 12) was related to a longer time to progression (HR 0.18, 95% CI 0.06-0.54) [48].
In light of recent studies showing beneficial activity by anti-HER2 drug combinations containing tucatinib or neratinib, and a lack of comparative data, it is impossible to assess the proper timing and place, if any, of the lapatinib/radiotherapy combination, or to state whether the novel TKIs would be more efficient in this context.

Challenges and Open Questions
The available data indicate remarkable intracranial efficacy from novel agents, such as tucatinib, trastuzumab deruxtecan and neratinib, even though lapatinib-containing combinations may still be relevant in this context. Treatment-induced diarrhea may be a limiting factor and requires proper surveillance and support. However, despite the major steps of the last decade and the cascade of new anti-HER2 agents that are being developed, some of them with remarkable intracranial efficacy, some questions still remain unanswered.
The guidelines for the management of advanced breast cancer do not support BM screening in patients with HER2-positive breast cancer. Nonetheless, as described in the current review, patients with de novo BM have different characteristics and better outcomes compared to those diagnosed with BM during the course of treatment [72]. In addition, there are, today, more treatment choices in HER2-positive BC with intracranial lesions than before, predisposing a higher probability of clinical benefit from early interventions. Hence, early detection may be crucial for patient survival and welfare. Hopefully, the ongoing NCT03881605, NCT03617341 and NCT04030507 trials will shed some light on this critical question.
On the other hand, the management of early HER2-positive BC has changed drastically during recent years, and there are currently no data to inform of the risk of BM among patients with pCR after neoadjuvant therapy or without pCR treated with adjuvant T-DM1. Similarly, it is unknown whether BM as a first site of distant recurrence presented after neoadjuvant double HER2-blockade, with or without subsequent adjuvant T-DM1, or after adjuvant dual blockade, will demonstrate the same response to the anti-HER2 therapies as the ones demonstrated thus far. All patients included in the DESTINY Breast01 trial had received prior trastuzumab and T-DM1, and about one third of the patients in the NALA study were previously treated with trastuzumab, pertuzumab and T-DM1 [26,35]. Thus, there is compelling evidence to suggest that these new agents have intracranial activity among pretreated patients, but the best timing for the administration of each agent and the optimal sequence remain uncertain.
In addition to elaborating on the preferred and more beneficial sequence of the various anti-HER2 agents, there is a great need for adjuvant studies investigating the role of the novel TKIs in preventing brain metastases. Both reduction in the incidence of BM and/or prolongation of time to development of BM would be of interest in this patient group.
In general, patients with BM have been excluded from clinical trials due to poor performance status, shorter expected survival and fears of increased toxicity [73]. Furthermore, investigators have been reluctant to include these patients due to technical obstacles such as limitations in the application of common assessment criteria such as RECIST or the need for additional imaging with different modalities. Additionally, matters related to patient history, such as concomitant corticosteroids and previous brain radiotherapy, may impede proper assessment in the context of a clinical trial [74]. Even more difficult is the evaluation of patients with a leptomeningeal disease, since the nature of the condition and the variety of symptoms hamper objective radiological size measures and reproducibility [75]. As evidence informing the intracranial activity of the new TKIs and ADC grows, clinical trials no longer exclude patients with BM by default, although enrollment of patients with progressive BM is still uncommon [76]. In the same spirit, the FDA issued a guidance for the industry last year, encouraging the inclusion of patients with BM, with some reservations for investigational drugs expected to have CNS toxicity (Guidance document, cancer clinical trial eligibility criteria: brain metastases, July 2020. Docket number: FDA-2019-D-0357, draft published 13 March 2019).
To extract information that could change practice in this area, it is not enough to only include patients with BM in the studies. Many trials do not have pre-specified efficacy outcomes or independent cohorts for this important subgroup, nor do they make a distinction between dissemination in the brain metastases and meninges. Although some trials assess intracranial activity, BM-related outcomes are usually not included as primary or co-primary endpoints. A summary of ongoing trials with endpoints related to BM is presented in Table 2. This notwithstanding, BM response criteria are different across various clinical trials, hampering comparison and conclusions. Standardization of procedures, endpoints and assessment tools is mandatory in order to facilitate critical appraisal of the studies and inference on the reported outcomes. With that aim, in 2015, the RANO-BM working group established recommendations for response evaluation in clinical trials [74]. The use of a common language should aid determination and discovery of clinical, biological and/or radiomic biomarkers predictive of CNS response, and enable tailoring of systemic and/or local treatments.
The heterogeneity of the clinical presentation of metastatic CNS disease also needs to be addressed. For example, leptomeningeal disease could also be present with or without parenchymal metastases in patients with BC; similarly, progression can refer to new lesions in the brain or in the meninges. Leptomeningeal disease can present with or without cancer cells in the cerebrospinal fluid, and is more uncommon than brain metastases in HER2-positive MBC [77,78]. Endpoints reporting on CNS disease outcomes do not usually distinguish between parenchymal metastases and leptomeningeal disease, thus providing insufficient data about the efficacy of anti-HER2 treatment in patients with leptomeningeal disease, either isolated or in combination with BM [5,77].   BM: brain metastases, CNS: central nervous system, DCR: disease control rate, MBC: metastatic breast cancer, MRI: magnetic resonance imaging, N/A: non-applicable, ORR: objective response rate, PFS: progression-free survival, RANO-BM: response assessment in neurooncology brain metastases. TNBC: triple-negative breast cancer.
Symptoms related to BM can be invalidating for the individual patient, which has special relevance in the context of improved survival in patients with HER2-positive BC. This dictates the need to address both disease-related symptoms and adverse events related to the treatment. It is important to report such aspects in a standardized manner in order to facilitate comparisons between the various agents and their impact on the global health of the patient [5]. For example, administration of neratinib in the TBCRC022 trial was accompanied with deterioration in quality of life, whereas no such relationship was observed in the NALA trial [34,35]. In this context, the use of the same standardized and validated questionnaire by both studies (EORTC QLQ-C30) allows for a more comprehensive comparison. TBCRC022 also demonstrated a decline in neurocognitive function of patients, another key parameter to be evaluated in the specific cohort [34]. Thus, inclusion of patient-reported outcomes, evaluation of neurocognitive function and quality of life aspects as endpoints in future trials are essential.
Taking into consideration the particularities of the CNS, alternative administration routes definitely merit formal investigation. Intrathecal trastuzumab is thought to have antitumoral activity in patients with leptomeningeal disease. A pooled analysis reported a response rate of 66.7%, with a median CNS-PFS of 7.5 months [79]. In another metaanalysis, clinical improvement was reported in 55% and stabilization in 14%, whereas median CNS-PFS was 5.2 months and median OS 13.2 was months [79]. Whether intrathecal administration of trastuzumab-pertuzumab or administration of ADC would provide better responses is far from explored, as is whether this administration route, with the inconvenience and the risks it involves, is more beneficial than the commonly used intravenous and oral routes.
Providing solid guidelines in this complex situation is difficult considering the heterogeneity of assessed outcomes, the backgrounds of included patients, the limited number of comparative studies and the disparities between countries in terms of drug availability and reimbursement of the novel drugs. The European Society for Medical Oncology (ESMO) guidelines, published last year, suggest considering the combination of tucatinib/trastuzumab/capecitabine in patients with HER2-positive BC with BM [5]. Although tucatinib is now approved by the FDA and the EMA, the drug is not globally available. Other combinations such as neratinib/capecitabine or lapatinib/capecitabine can also be of value in this setting, and results from novel drugs, such as trastuzumab deruxtecan, have thus far been promising and could provide more treatment options. The ideal sequence for using all the available agents is unclear at this time, but it is generally accepted that exposure to the different available anti-HER2 agents may improve patients' outcomes and should be pursued whenever possible.

Conclusions
BM is a major cause of morbidity and mortality in patients with MBC [1]. To date, local management, either ablative (surgery) or palliative (radiotherapy), has been the preferred option in the management of BM. On the other hand, these treatments are not exempt from side effects, which may further deteriorate new neurological function [60]. Aside from anti-HER2 monoclonal antibodies, TKIs and ADCs have exhibited antineoplastic intracranial activity, either as a response to established CNS metastases or by delaying time in the development of BM. These treatments may substitute, delay or complement local treatment strategies in order to attain the best control of CNS disease. Novel anti-HER2 therapies, clinical trials with CNS disease-specific cohorts, standardization of endpoints and evaluation criteria (both in prevention and in established BM), as well as screening for BM at baseline, are fundamental elements in order to improve management, symptom control and prognosis of HER2-positive MBC patients.